Clustering forest harvest stands on spatial networks for optimised harvest scheduling

Context

Clustering forest harvest stands reduces the movements of the harvesters, forwarders and staff. Moreover, it simplifies the subsequent log transport, when compared with dispersed stands.

Aims

Harvesting activities are generally based on silvicultural motivated planning data. The development of an analytical method to cluster harvest stands with respect to the spatial network of roads should improve the harvesting effort.

Material and method

The clustering of harvest stands was developed for Aracruz (Brazil) in 2004, where it is used there successfully since. The hierarchical method ??single linkage cluster analysis?? is applied. As a distance function, the Euclidian distance was substituted by the shortest path on the spatial network.

Result

The clustering method is based on the minimum spanning tree, which is the spatial equivalent to the dendrogram of an ordinary cluster analysis. Applying the Delaunay triangulation to fill the distance matrix reduces the distance calculation time from O(n ²) to O(n). The method is illustrated by a planning district of the Aracruz enterprise.

Conclusion

Harvesting units are properly clustered spatially by the discussed method. Topographic obstacles are automatically avoided and the need to relocate machinery is reduced as well as the total driving distance.